This article is written by students.

It may include omissions and imperfections,

as far as possible, reported by our reviewers in the editorial notes.

A dice game
[Year 2019 – 2020]

SURNAMES AND FIRST NAMES OF STUDENTS, GRADES: Sabrina-Elena BENCHEA,

George-Octavian GRUMĂZESCU, Mara-Elena MARDARASEVICI, Ioana MUHA, Tudor-
Andrei PRICOP, George-Constantin RUSU, George-Ioan STOICA, students in 10th grade.

SCHOOL: Colegiul Naţional Costache Negruzzi, Iaşi

TEACHER: Adrian Zanoschi

RESEARCHER: Iulian Stoleriu, Faculty of Mathematics, “Alexandru Ioan Cuza” University

Iași.

1. PRESENTATION OF THE RESEARCH TOPIC

The paper deals with the winning chances in a game of dice. We firstly consider the distribution
of the sum of a set of four classical dice that are rolled together, then we consider other types of
dice, which take the form of the remaining four Platonic solids.

2. BRIEF PRESENTATION OF THE CONJECTURES AND RESULTS OBTAINED

At a certain casino, four fair classical dice are thrown simultaneously. If a player bets on a
certain sum, what are the chances of winning? Find the most probable sum of the points that
appear on the top faces and the expected number of throws needed to obtain it for the first time?
You may also consider similar questions for the case of four identical dice that take the
form of a general regular convex polyhedron.

MATh.en.JEANS 2019-2020 [Colegiul Național Costache Negruzzi, Iași] page 1

 This article is written by students. It may include omissions and imperfections,
as far as possible, reported by our reviewers in the editorial notes.

THE SOLUTION

At a certain casino, four fair dice are thrown simultaneously. If a player bets on a certain
sum, what are the chances of winning? What is the most probable sum and what is the
expected number of throws needed to obtain it for the first time?

Method I.

Let’s suppose that the numbers shown on the four dice after one throw are x1 , x2 , x3 , x4 ,

where xi  {1, 2, 3, 4, 5, 6} , i  1, 4 .

Denote the sum of these values by S k  k  x1  x2  x3  x4 and the number of possible

cases as Pk . We observe that S k  {4, 5, 6, ..., 23, 24} . Thus, it can be attributed 21 values. We
will prove that the sum with the highest chance of appearance is S14  14 which means that P14 is

the greatest out of the numbers Pk , k  4, 24 .

Because the equation:

x1  x2  x3  x4  k , whilst xi  {1, 2, 3, 4, 5, 6} , i  1, 4

is equivalent to the equation:

(7  x1 )  (7  x2 )  (7  x3 )  (7  x4 )  28  k , whilst 7  xi  {1, 2, 3, 4, 5, 6} , i  1, 4 ,

we have Pk  P28 k , for any k  {4, 5, ..., 13} .

MATh.en.JEANS 2019-2020 [Colegiul Național Costache Negruzzi, Iași] page 2
 This article is written by students. It may include omissions and imperfections,
as far as possible, reported by our reviewers in the editorial notes.

Thus, it suffices to determine P4 , P5 , ..., P13 and P14 . To do this, we shall prove the following
lemma:

Lemma. The number of solutions of the equation:

x1  x 2  x3  x4  k ,
(1)

 k  1
 
where x1 , x2 , x3 , x4 , k are positive integers and k  4 is equal to  3  .

Proof. Any solution of equation (1) can be obtained by placing the 3 sticks between the terms of
a string of k ones (between two terms we can place no more than one stick). For example, the
figure:

1.1.1
 1.1.1.1.1
   1.1.1.1.1
   1
x1  3 x2  5 x 1 x3  5 4

represents the solution x1  3, x2  5, x3  5, x4  1 of the equation x 1  x2  x3  x4  14 .

The number of solutions of equation (1) is equal to the number of ways in which we can
choose three places to place the three sticks out of k  1 spaces between the terms of the string.
 k  1
 
Thus, the result is  3  .

Let’s return to the initial idea of finding the number of possibilities P4 , P5 , ..., P13 and P14 .

 k  1
 
According to the Lemma, equation (1) has  3  solutions for any k  {4, 5, ..., 13, 14} .

But because x1 , x2 , x3 , x4 are numbers written on dice, we have xi  {1, 2, 3, 4, 5, 6} , i  1, 4 , so

 k  1
 
some of the  3  solutions will not work.

If ( x1 , x2 , x3 , x4 ) is a solution of equation (1) and at least 2 of its components are greater

than 6, then:

x1  x2  x3  x4  1  1  7  7  16

MATh.en.JEANS 2019-2020 [Colegiul Național Costache Negruzzi, Iași] page 3

 This article is written by students. It may include omissions and imperfections,
as far as possible, reported by our reviewers in the editorial notes.

Because k  14 , we know that at most one of the components of the solution ( x1 , x2 , x3 , x4 )

could be greater than 6.

k   4, 5, 6, 7, 8, 9
For , all the numbers x1 , x2 , x3 , x4 are less than or equal to 6, therefore, in each
 k  1  4  1

3
 P   3   1,
one of the 6 cases, the number of solutions for equation (1) is   , so 4  

 5  1  6  1  7  1  8  1  9  1
P5     4, P6     10, P7     20, P8     35, P9     56.
 3   3   3   3   3 

For k  10 , equation (1) has, except for the valid solutions, the solutions: (7, 1, 1, 1), (1,7,1,1),
 10  1
P10     4  80
(1,1,7,1), (1,1,1,7). So,  3  .

For k  11 , equation (1) has, other than the valid solutions, the solutions: (8, 1, 1, 1) with its
 11  1
P11   
permutations (4 in total) and (7, 2, 1, 1) with its permutations (12 in total). So,  3 
(4  12)  104 .

For k  12 , equation (1) has the following non-valid solutions: (9, 1, 1, 1) and its permutations (4
in total), (8, 2, 1 ,1) and its permutations (12 in total), (7, 3, 1, 1) and its permutations (12 in total)
 12  1 
P12   
and (7, 2, 2, 1) and its permutations (12 in total). Thus,  3  (4  12  12  12)  125 .

For k  13 , equation (1) has, other than the valid solutions, the solutions: (10, 1, 1, 1) and its
permutations (4 in total), (9, 2, 1 ,1) and its permutations (12 in total), (8, 3, 1, 1) and its
permutations (12 in total), (8, 2, 2, 1) and its permutations (12 in total), (7, 4, 1, 1) and its
permutations (12 in total), (7, 3, 2, 1), and its permutations (24 in total) and (7, 2, 2, 2) and its
 13  1
P13     (4  12  12  12  12  24  4)  140
permutations (4 in total). Hence,  3  .

For k  14 , equation (1) has the following non-valid solutions: (11, 1, 1, 1) and its permutations
(4 in total), (10, 2, 1 ,1) and its permutations (12 in total), (9, 3, 1, 1) and its permutations (12 in
total), (9, 2, 2, 1) and its permutations (12 in total), (8, 4, 1, 1) and its permutations (12 in total),
(8, 3, 2, 1), and its permutations (24 in total), (8, 2, 2, 2) and its permutations (4 in total), (7, 5, 1,

MATh.en.JEANS 2019-2020 [Colegiul Național Costache Negruzzi, Iași] page 4

 This article is written by students. It may include omissions and imperfections,
as far as possible, reported by our reviewers in the editorial notes.

1) and its permutations (12 in total), (7, 4, 2, 1) and its permutations (24 in total), (7, 3, 3, 1) and
 14  1
P14   
its permutations (12 in total), and (7, 3, 2, 2) and its permutations (12 in total), so  3 
(4  12  12  12  12  24  4  12  24  12  12)  146 .

We organise all the results in the following table:

Sk 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Pk 1 4 10 20 35 56 80 104 125 140 146 140 125 104 80 56 35 20 10 4 1

Thus, the sum with the highest chance of appearing is S14  14 .

Method II.

Due to the fact that 6  1296 is a very small number (for a computer), we can take all the
4

possibilities and count the sum at each step.

We have written 21 lines, each line showing the sum and the probability of that sum to appear.

MATh.en.JEANS 2019-2020 [Colegiul Național Costache Negruzzi, Iași] page 5

 This article is written by students. It may include omissions and imperfections,
as far as possible, reported by our reviewers in the editorial notes.

As we can see above, the sum S14  14 has the biggest chance of appearing. The probability of
obtaining this sum is equal to:

number of favorable cases 146 146 73

p  4  
number of possbile cases 6 1296 648

Let q  1  p .

Denote the number of necessary tries for obtaining sum 14 for the first time by the discrete
random variable X and with pk the probability of obtaining sum 14 for the first time after throw
number k. If we don’t succeed after k  1 tries and we do succeed after the next throw then,
k 1
evidently, the first time we’ll get the sum 14 is after throw number k. So, pk  q  p .

We have:

X 1 2 3 4 … k …
Prob. p1  p p2  q  p p3  q 2  p p4  q 3  p … pk  q k 1  p …

The expected number of throws of the four dice until we observe sum 14 is:
n
E  X   1  p  2  q  p  3  q 2  p  ...  k  q k 1  p  ...  lim  k  q k 1  p
n 
k 1 .

Now, we need to compute the sum:

n
En   k  q k 1  p
k 1 .

We have:

En  q  En  p  2  q  p  3  q 2  p  ...  n  q n 1  p  q  p  2  q 2  p  ...  (n  1)  q n 1  p  n  q n  p

 p  q  p  q 2  p  ...  q n 1  p  n  q n  p
1  qn p  (1  q n )  n  q n  p  (1  q )
 p  n  qn  p  ,
1 q 1 q

from which we deduce that:

MATh.en.JEANS 2019-2020 [Colegiul Național Costache Negruzzi, Iași] page 6

 This article is written by students. It may include omissions and imperfections,
as far as possible, reported by our reviewers in the editorial notes.

p  p  (1  n)  q n  n  q n 1
En 
(1  q) 2 .

q   0,1
Because , we have:

lim (1  n )  q n  lim n  q n 1  0
n  n  ,
n
p p 1
E ( X )  lim
n 
k  q
k 1
k 1
 p  lim En 
n (1  q) 2
 2
p p.

Thus, the expected number of throws of four dice until we first get sum 14 is:

1 648
  8,87  9
p 73 .

In order to find out the E  X  we can also use this formula:

E  X   E  X for success in the first throw   p  E  X for failure in the first throw    1  p 
 1  p   E  X   1   1  p   1   1  p  E  X 
,
1
E X  
so E  X    1  p  E  X   1 , therefore p.

It is known for more than two thousands years that the only regular and convex polyhedra are the
Platonic solids. These polyhedra are named after the ancient Greek philosopher Plato. Except for
cube (regular hexahedron), they are:

1. regular tetrahedron
2. regular octahedron
3. regular dodecahedron
4. regular icosahedron

We now want to solve the same problem when the dice would have each of the above shapes.

1. If the dice are shaped as a regular tetrahedron, then the sum S k  {4, 5, 6, ..., 15,16} . Using the
same reasoning as with the regular cubic dice, we get the following table:

Sk 4 5 6 7 8 9 10 11 12 13 14 15 16
Pk 1 4 10 20 31 40 44 40 31 20 10 4 1

MATh.en.JEANS 2019-2020 [Colegiul Național Costache Negruzzi, Iași] page 7

 This article is written by students. It may include omissions and imperfections,
as far as possible, reported by our reviewers in the editorial notes.

Thus, the sum with the highest chance of appearing is S10  10 and the expected number of
throws of four dice until we first get sum 10 is:

1 64
  5,81  6
p 11 .

2. If the dice are shaped as a regular octahedron, then the sum S k  {4, 5, 6, ..., 31,32} . Using the
same reasoning as with the regular cubic dice, we get the following table:

Sk 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Pk 1 4 10 20 35 56 84 120 161 204 246 284 315 336

Sk 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
Pk 344 336 315 284 246 204 161 120 84 56 35 20 10 4 1

Thus, the sum with the highest chance of appearing is S18  18 and the expected number of
throws of four dice until we first get sum 18 is:

1 84
  11,9  12
p 344 .

3. If the dice are shaped as a regular dodecahedron, then the sum Sk  {4, 5, 6, ..., 47, 48} . Using
the same reasoning as with the regular cubic dice, we get the following table:

Sk 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Pk 1 4 10 20 35 56 84 12 165 22 286 36 451 544 64 736
0 0 4 0

Sk 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Pk 82 91 99 106 111 114 115 114 111 106 99 91 82 73 640
9 6 4 0 1 4 6 4 1 0 4 6 9 6

Sk 35 36 37 38 39 40 41 42 43 44 45 46 47 48
Pk 544 451 364 286 220 165 120 84 56 35 20 10 4 1

MATh.en.JEANS 2019-2020 [Colegiul Național Costache Negruzzi, Iași] page 8

 This article is written by students. It may include omissions and imperfections,
as far as possible, reported by our reviewers in the editorial notes.

Thus, the sum with the highest chance of appearing is S 26  26 and the expected number of
throws of four dice until we first get sum 26 is:

1 124
  17,93  18
p 1156 .

4. In the case of a regular icosahedron, we can see, by using a computer, that the sum with the
highest chance of appearing is S 42  42 and the expected number of throws of four dice until we
first get sum 42 is:

1 204
  29,96  30
p 5340 .

3. CONCLUSION

In this paper, we have found the distribution of the sum of points that appear on the top faces
when four classical identical dice are rolled together. Also, we have considered a similar problem
when the dice take the form of a: regular tetrahedron, regular octahedron, regular dodecahedron
and regular icosahedron. Together with the regular hexahedron (the cube), these five polyhedra
are the only regular convex polyhedra, also called Platonic solids.

In each case, we have calculated the expected number of throws until the most probable sum
appears for the first time.

MATh.en.JEANS 2019-2020 [Colegiul Național Costache Negruzzi, Iași] page 9